1. Field of the Invention
The present invention relates to an image quality correction circuit for correcting the quality of the images produced by video equipment, such as television receivers, video cameras, and the like.
2. Description of Related Art
In the NTSC, PAL, and SECAM television signal transmission systems, wide-band three primary color signals, R, G, and B, are first subjected to gamma correction, which is necessary to compensate for the equipment characteristic at the receiving end, and are then converted into a luminance signal Y and color-difference signals, R-Y and B-Y, or chrominance signals, I and Q, for transmission, the bandwidths being limited to about 0.5 to 1.5 MHz for the color-difference signals or chrominance signals.
The gamma correction and color-difference signal bandwidth limiting performed at tile transmitting end, however, result in the introduction of nonlinearity in the transmitted signal, and in the case of a high color saturation image, high-frequency components contained therein cannot be reproduced satisfactorily which normally should be reproduced completely by the luminance signal alone. In other words, the high-frequency component level of the luminance signal degrades in the high color saturation areas of the image, and fine details of the original scene cannot be displayed sufficiently. It is also known that black level variations axed saturation drops occur in the high-frequency component areas of a high color saturation image.
Image quality correction circuits designed to prevent such image quality degradations include one such as disclosed in Japanese Patent Application Laid-Open No. 64-32588 (1989). FIG. 1 shows a block diagram for the image quality correction circuit disclosed therein. The luminance signal Y is inputted to a high-pass filter 63 and also to a delay circuit 65. The high-frequency component of the luminance signal passed through the high-pass filter 63 is fed to a variable gain amplifier 64 whose output signal is supplied to an adder 66. Also inputted to the adder 66 is the luminance signal Y delayed through the delay circuit 65. The adder 66 adds together the high-frequency component of the luminance signal fed from the variable gain amplifier 64 and the luminance signal Y delayed through the delay circuit 65, and outputs a corrected luminance signal Y'.
On the other hand, the color-difference signal R-Y is inputted to a full-wave rectifier 10, and the full-wave rectified color difference signal R-Y is inputted to an adder 13. Similarly, the color-difference signal B-Y is inputted to a full-wave rectifier 12, and the full-wave rectified color-difference signal B-Y is inputted to the adder 13, where the color-difference signal R-Y and the color-difference signal B-Y are added together. The color density thus detected is outputted as a color density detection signal which is applied to a control terminal of the variable gain amplifier 64.
The operation of the above image quality correction circuit will be described below. The luminance signal Y is inputted to the high-pass filter 63, where the high-frequency component of the luminance signal Y is separated and inputted to the variable gain amplifier 64. The gain of the variable gain amplifier 64 is controlled in accordance with the color density detection signal outputted from the adder 13 by detecting the color density. More specifically, in high color-density areas, the amplitude of the color density detection signal is increased, so that the gain with which to amplify the high-frequency component of the luminance signal is increased; conversely, in low color-density areas, the amplitude of the color density detection signal is reduced, so that the gain with which to amplify the high-frequency component is reduced.
The delay circuit 65 delays the luminance signal Y before inputting to the adder 66, so that the phase of the luminance signal Y to be corrected coincides with the phase of an image quality correction signal, that is, the output of the variable gain amplifier 64, representing the high-frequency component of the luminance signal. The adder 66 adds together the luminance signal fed from the delay circuit 65 and the image quality correction signal fed from the variable gain amplifier 64, and outputs the corrected luminance signal Y'. Thus, the luminance signal, outputted as the luminance signal Y', is so corrected that the gain of the high-frequency component of the luminance signal is increased in the high color-density areas.
The color-difference signal R-Y is full-wave rectified by the full-wave rectifier circuit 10; when we consider a vector of the color-difference signal R-Y, the amplitude of the rectifier output represents the length of the R-Y vector. The color-difference signal B-Y is full-wave rectified by the full-wave rectifier circuit 12, and the amplitude of its output represents the length of the color-difference signal B-Y vector. The full-wave rectified color-difference signal R-Y outputted from the full-wave rectifier circuit 10 and the full-wave rectified color-difference signal B-Y outputted from the full-wave rectifier circuit 12 are added together in the adder 13. Although the output signal of the adder 13 does not become equal to the length of the resultant of the color-difference signal R-Y and B-Y vectors, the output signal can be regarded, for simplicity, as representing the color density; the color-difference signals have a larger amplitude in the high color-density areas and a smaller amplitude in the low color-density areas. The gain of the variable gain amplifier 64 is controlled in accordance with the color density detection signal outputted from the adder 13 by detecting the color density.
As described above, according to the prior art image quality correction circuit, when the amount of correction is increased for the high-frequency component of the luminance signal, the amplitude of the high-frequency component of the luminance signal increases in the high color-density areas, but in areas where the high-frequency component of the luminance signal is large in the positive side, the amplitude of the luminance signal alone increases while the magnitude of the color signals does not increase. This causes color dropout, resulting in image quality degradation. That is, a saturation drop is exacerbated in the high-frequency areas of a high color saturation image. This tendency is particularly pronounced in areas where over-shoots and preshoots occur. Furthermore, in the high color-density areas, the signal-to-noise ratio decreases since the noise component of the luminance signal is also amplified.
That is, as the amount of correction is increased in the high color saturation areas, such problems as poor detail reproduction and black level variation can be alleviated correspondingly, but this in turn causes the problem of increased saturation drop and S/N degradation, which places a limit on the amount of correction that can be achieved. Therefore, the image quality improvement that can be perceived by the eye has not been satisfactory.
Furthermore, in the case of the color density detection signal produced in the prior art image quality correction circuit, the amount of correction is not distributed appropriately between various colors. As previously noted, in the NTSC, PAL, and SECAM television signal systems, the high-frequency component of the video signal is transmitted by the luminance signal alone, and since the amplitude ratio of the luminance signal contained in each color is different, the amount of high-frequency component reduction is also different. That is, the amount of high-frequency component reduction is small in areas of a color containing a large amount of luminance component, while the amount of high-frequency component reduction is large in areas of a color containing a small amount of luminance component. In a specific example, for an image of monochromatic blue consisting of 100% B signal, the amplitude of the luminance signal can be calculated as Y=0.11 from the equation Y=0.30R+0.59G+0.11 B since R=G=0 and B=1. The value is the smallest of all the colors in the color bars. It is proven that, in this case, the gain of the high-frequency component is reduced to 11% of the gain before transmission at the transmitting end, supposing that .gamma. characteristic of the television picture tube is 2.0. Accordingly, the gain of the high-frequency component drops down to the amplitude ratio of the luminance signal. If high frequency components are contained in the colors of the color bars, the high-frequency component in each color drops down to the ratio shown in Table 1 below.
TABLE 1 ______________________________________ Red Green Blue Magenta Cyan Yellow White ______________________________________ 30% 59% 11% 41% 70% 89% 100% ______________________________________
Each of the values shown in Table 1 coincides with the amplitude ratio of the luminance signal contained in each color. If the high-frequency component is to be corrected for each color, the amplitude ratio, 11%, of the luminance signal for the blue color, for example, requires that the high-frequency component should be corrected to the ratio of 1/0.11. In the prior art example, no consideration is given to the amplitude ratio of the luminance signal contained in each color. If this factor is to be considered, the color density signal obtained in the prior art example needs to be divided by tile luminance signal.
There is disclosed another prior art which proposes an example involving division by the luminance signal, but one shortcoming of this example is that the complexity of circuitry increases because of the inclusion of a dividing circuit in the electric circuit.
Another problem is the effect of correction appearing unnatural at the boundaries between colors. FIG. 2 shows how a primary color signal is affected when a color-difference signal is created from the luminance signal and when the high-frequency component of the luminance signal is enhanced, by taking as an example a pattern consisting of successive color bars, i.e., gray, red, white, red, and black arranged in this order from left, to right on the screen. Solid line 170 indicates the luminance signal; dotted lines show the contours of the portions where high-frequency enhancement are made; 171, 175 indicate the black level; 172 is the R-Y color-difference signal; 173 is a no-color level; solid line 174 represents the R primary color signal; and dotted lines show the portions where the high-frequency correction is made to the luminance signal 170, the reference signals A, B, C, and D from left to right indicating the waveforms at the respective color boundaries. As can be seen from FIG. 2, high-frequency correction is effective in achieving uniform image quality only in the case of D, but in the cases of A, B, and C, the high-frequency correction of the luminance signal causes unnatural contours. A, B, and C are where the slope is reversed between the luminance signal 170 and the color-difference signal 172. In A, high-frequency correction results in an unnatural step formed in the rising portion of the primary color signal. In B and C, the boundary contours which initially were not present on the reverse image side are formed because the luminance and color-difference signals are transmitted separately in separate frequency bands. These contours are further emphasized by the high-frequency correction of the luminance signal. As a result, overcorrection tends to occur in the case of the B and C patterns, resulting in overemphasized contours.
There are other problems: in the case of an image whose overall S/N ratio is low, if correction is made meticulously on light color portions, the S/N ratio will further degrade, and furthermore, while the appearance of wrinkles in the human skin, a light color area, should be reduced to obtain a pleasing image, if correction is made to such light color areas, the image will appear more real, making the wrinkles further noticeable.
A further problem is that aperture correction is performed using a separate circuit, requiring the provision of a separate aperture correction circuit and thus increasing the size of the circuitry required to achieve image improvements.